Exploratory Data Analysis (EDA) Projects in Python

Published:

Problem

Data scientists often encounter diverse datasets requiring tailored cleaning, transformation, and exploratory techniques before modelling.
Goal: Build a reproducible Python EDA framework demonstrating how to extract insights, engineer features, and communicate patterns from unstructured datasets across different domains — public safety (911 calls) and financial markets.

Approach

  • Designed two end-to-end EDA pipelines using real-world datasets:
    1. 911 Calls Analysis: time and location-based patterns of emergency calls.
    2. Finance Data Analysis: stock price behaviour, returns, and inter-company correlations.
  • Implemented data ingestion → cleaning → transformation → visualisation using pandas and numpy for data handling and seaborn/plotly for insight communication.
  • Created reusable analysis templates for:
    • Date/time feature engineering (.dt accessors, grouping, resampling)
    • String and categorical handling (type conversion, feature splitting)
    • Correlation and pairwise analysis
    • Multi-panel and interactive visualisations for pattern discovery.

Stack

  • Language: Python 3
  • Libraries: pandas, numpy, matplotlib, seaborn, plotly, datetime
  • Tools: Jupyter Notebook, Git/GitHub
  • Concepts: EDA, data cleaning, feature extraction, time series analysis, correlation analysis, visualisation design

Case Studies

1. 911 Calls EDA

Objective: Explore temporal and spatial patterns in emergency call data.

  • Parsed timestamps into year, month, day, and hour features for time-based analysis.
  • Mapped call reasons and types to broader categories (e.g., EMS, Fire, Traffic).
  • Visualised daily and monthly call volume, call-type distributions, and temporal trends.
  • Identified operational peaks and seasonal call variation patterns.

Skills: datetime manipulation, grouping and aggregation, categorical encoding, visualisation (line, bar, count, heatmap).

2. Finance Data EDA

Objective: Investigate stock price dynamics and inter-company behaviour.

  • Collected multi-stock price data via Yahoo Finance API.
  • Calculated moving averages, daily returns, and cumulative returns.
  • Conducted pairwise correlation and risk–return analysis across multiple tickers.
  • Visualised price trends and co-movement patterns through heatmaps and scatter matrices.

Skills: time-series analysis, rolling windows, correlation matrices, multi-plot visual storytelling.

Results

  • Demonstrated consistent EDA methodology applicable across domains.
  • Built a reproducible framework highlighting how to structure exploratory workflows for both categorical–temporal and continuous–financial data.
  • Strengthened proficiency in data storytelling and visualisation using modern Python tools.

Impact

  • Forms the analytical bridge between raw data handling and predictive modelling.
  • Provides an adaptable template for future projects involving data cleaning and insight extraction.
  • Complements the “Python OOP Mini-Systems” repository by demonstrating data-centric rather than logic-centric Python application.

Links & Resources